- Shell 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
For a quick standalone rerun to verify the memcpy host-vs-default gap observed in the full cpu-type-comparison sweep, without re-running the whole 12-test suite. |
||
| schema | ||
| test-configs | ||
| benchmarks.json | ||
| LICENSE | ||
| README.md | ||
| run-pts.sh | ||
| setup-pts.sh | ||
Test Addressed PTS Tooling
Scripts for running standardized benchmarks for Tech Addressed articles. Uses the Phoronix Test Suite.
Files
setup-pts.sh— Installs prerequisites, clones PTS, and installs the tests defined in a test config filerun-pts.sh— Runs a benchmark test session and writes results to JSONbenchmarks.json— Master registry of all benchmark tests, versions, and result definitions used by Tech Addressedtest-configs/— Test configuration files (one per testing scenario)schema/*.schema.json— JSON schema files for validating data files
Usage
1. Setup
Run once on a new system before benchmarking:
./setup-pts.sh <test-config>
Test config files are located in test-configs/. Example:
./setup-pts.sh standard-tests.json
This will update the system, install dependencies, clone PTS, and install all tests defined in the config.
2. Run benchmark tests
./run-pts.sh <test-config> <device-slug> <config-name> <temp-device-num> [session-id]
| Argument | Description |
|---|---|
test-config |
Test config filename (e.g. standard-tests.json) |
device-slug |
Short identifier for the device (e.g. rock-5b) |
config-name |
Testing configuration name (e.g. stock) — determines the result filename |
temp-device-num |
Position of the CPU/SoC temperature block in sensors output (1-based) |
session-id |
(optional) Integer session ID. Auto-incremented if omitted. |
Example:
./run-pts.sh standard-tests.json inovato-quadra stock 2
Results are written to ~/benchmarks/<device-slug>/<config-name>.json.
Finding the correct temp-device-num: Run sensors on the device and count the adapter blocks from the top. Pass the number of the block that represents the CPU/SoC package temperature.
Key Concepts
Session ID — All results from a single benchmark session share the same integer session_id. This allows results to be grouped. A value of 0 indicates a standalone individual test run. The session ID is auto-incremented from the last run if not specified.
Test Versions — The version field in benchmarks.json is the approved PTS version for each test. All systems should be benchmarked using the same version to ensure comparable results. When a version is updated, the old entry is retired and a new one is added.
Retired Tests — Tests are retired rather than deleted. Retired tests are excluded from new runs but their historical results remain valid.